participation in the design of performance management ... · the speciﬁc form of participation...

Journal of Organizational Behavior

J. Organiz. Behav. 25, 831–851 (2004)

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/job.266

Participation in the design of performancemanagement systems: a quasi-experimentalfield study

AD KLEINGELD*, HARRIE VAN TUIJL AND JEN A. ALGERA

Department of Technology Management, Technische Universiteit Eindhoven, Eindhoven, The Netherlands

Summary In the literature on the relationship between participation in decision making and perfor-mance, a tell-and-sell strategy is considered a viable alternative to participation. In contrast,we argue that in organizational settings, when a sensitive and important issue is at stake, par-ticipation of a form to be characterized as formal, long term, direct, and with a high degree ofparticipant influence is more effective than a tell-and-sell strategy. Using a quasi-experimentaldesign with a participation, a tell-and-sell, and a control condition, a ProMES performancemanagement system was implemented in the field service department of a Dutch supplierof photocopiers. Outcome feedback to individual technicians resulted in an average perfor-mance increase in the participation condition that was significantly higher than the increasefound in the tell-and-sell condition. Satisfaction with the program, and the perceived useful-ness of the feedback, were significantly higher in the participation condition. In both experi-mental conditions, the performance increase was significant compared to the controlcondition. An explanation for these findings is discussed, as are implications for theory andpractice. Copyright # 2004 John Wiley & Sons, Ltd.

Introduction

Methods for designing performance management systems (PMS), such as the productivity measure-

ment and enhancement system ProMES (e.g., Pritchard, 1990, 1995a; Pritchard, Holling, Lammers, &

Clark, 2002) or Management by Objectives (e.g., Drucker, 1976; Rodgers & Hunter, 1991), emphasize

the importance of employee participation in system design. It has been convincingly demonstrated that

participatively designed performance management systems can indeed result in significant and sub-

stantial performance increases (Pritchard et al., 1988, 1989). However, developing and implementing

performance management systems in real-life organizations entails a large number of interventions in

addition to participation. The design of a PMS entails components such as performance indicators,

goals, and feedback. Implementing a PMS implies, among others, providing feedback, discussing

feedback, solving problems, and setting goals. Thus, the question arises whether the participative

Received 15 April 2002Copyright # 2004 John Wiley & Sons, Ltd. Accepted 14 April 2004

* Correspondence to: Ad Kleingeld, Department of Technology Management, Technische Universiteit Eindhoven, 5600 MB,Eindhoven, The Netherlands. E-mail: [email protected]

development of a PMS, being only one of several interventions involved, indeed plays the critical role

that often is assumed. The purpose of this article is to shed some light on this issue. More precisely, the

central question is whether it is possible to obtain performance increases of the magnitude reported in

the literature, when, instead of implementing a participatively developed PMS, an identical PMS is

implemented, preceded by a tell-and-sell introduction of that system.

ProMES: productivity measurement and enhancement system

In order to be as precise as possible about the context in which we will address the above question and

about form of participation as operationalized in this study, we will start with a short introduction of

ProMES. ProMES is a participative development method for performance management systems, the

design of which has to fulfill a number of specific design requirements. Participative development

entails the involvement of employees in generating, evaluating, and deciding on all components of

the PMS that they themselves will use as a control loop for self-regulation. These components are:

the employees’ areas of responsibility, performance indicators (operational definitions of relevant

aspects of those areas of responsibility), transformation functions through which the contribution of

indicator performance to overall performance can be determined (so-called contingencies), and a feed-

back report. The development process, in which employees and their direct supervisor take part, is

guided by a facilitator who ensures that all participants adhere to a core decision-making rule—agreed

upon beforehand—that implies that decisions are made on a basis of consensus arrived at through a

process of discussion in terms of arguments. An important aid in judging the quality of arguments is a

number of evaluation criteria or design requirements. For example, indicators should be controllable,

i.e., employees should be the main factor of influence on indicators included in the system. Thus, argu-

ments for and against the controllability of proposed indicators are thoroughly discussed in develop-

ment sessions. As soon as the development team has reached consensus on the basic components of the

PMS, there is a next round of participative decision making in which those components are proposed to

management in one or more review and approval sessions. These sessions are also guided by the facil-

itator who sees to it that all participants adhere to the agreed decision-making rule of discussion in

terms of arguments, until consensus is reached. Very often, consensus is only reached after some nego-

tiation has taken place. For example, in many cases the development team is only prepared to take

responsibility for a particular performance indicator, if management allows for the resources or auton-

omy necessary to make the indicator controllable by the employees.

Implementing a performance management system that has beendesigned participatively versus assigned by means of tell-and-sell

Typical components of a PMS such as developed by ProMES are multidimensional performance indi-

cators, feedback, and objectives or goals. The positive effects of combining specific and challenging

goals with timely, specific, and positive outcome feedback have been well documented, both in labora-

tory and in field settings (e.g., Algera, 1990; Alvero, Bucklin, & Austin, 2001; Balcazar, Hopkins, &

Suarez, 1986; Kluger & DeNisi, 1996; Locke & Latham, 1990, 2002). Given the effectiveness of such

interventions, the question to be addressed in this study is whether there is added value in participa-

tively developing the components of PMS that enable such interventions. Methods for designing PMS

emphasize the importance of employee participation in system design, assuming that high participa-

tion produces cognitive benefits (e.g., better understanding of job priorities, development of effective

task strategies) and motivational gains (e.g., commitment to the system, acceptance of feedback and

832 A. KLEINGELD ET AL.

Copyright # 2004 John Wiley & Sons, Ltd. J. Organiz. Behav. 25, 831–851 (2004)

goals) (e.g., Pritchard, 1990; Rodgers & Hunter, 1991). The same arguments are mentioned in the lit-

erature on participation (Latham, Winters, & Locke, 1994; Locke & Schweiger, 1979; Wagner, Leana,

Locke, & Schweiger, 1997) when it comes to the assumed mechanisms underlying the effects of

participation: performance increases due to participation are attributed to both cognitive mechanisms

(such as increased communication, better utilization of knowledge, increased understanding of the job)

and motivational mechanisms (less resistance to change, sense of control, growth in mutual

trust, acceptance of and commitment to changes and decisions). Actually, these arguments go back

as far as the classical studies on group decision making by Coch and French (1948), Lewin (1952),

and Maier (1973).

We join this line of reasoning and hypothesize that implementation of a PMS that is developed

participatively in accordance with ProMES guidelines will have larger effects on performance com-

pared to the implementation of an identical PMS that is introduced by means of a tell-and-sell intro-

duction, because participation operationalized as above will have larger positive effects on the

cognitive and motivational mechanisms mentioned than a tell-and-sell introduction. This can be illu-

strated as follows.

The specific form of participation focused on in this study, namely the type of participation used in

ProMES, puts a heavy emphasis on joint discussions by employees and management on the controll-

ability of performance indicators. This careful review of arguments for and against controllability of

performance indicators can be expected to have an effect on performance through the cognitive

mechanisms mentioned above. More in particular, these discussions will result in clear-cut ideas on

what employees have to do. In addition, these discussions will result in better insights in how they can

do what they have to do. In Naylor, Pritchard, and Ilgen’s (1980) theory of behavior in organizations,

on which the ProMES method is based (Pritchard, 1995b), motivation is conceived of as the allocation

of time and effort to tasks. Consequently, knowing what to do and how to do it are important precondi-

tions for motivation. For instance, by informing one another on the existence of effective task strate-

gies, employees become aware of controllable routes to task completion. This will probably enhance

feelings of self-efficacy (Bandura, 1997) and increase expectancies (Vroom, 1964). A positive effect of

participation in formulating task strategies on performance, mediated by both self-efficacy and quality

of task strategies, was indeed demonstrated by Latham et al. (1994), who state that ‘People who know

that they have effective ways of performing a task will be more confident than people who are unsure

of how to perform effectively’ (Latham et al., 1994, p. 61).

A second important element of the form of participation employed in this study is the bottom-up

procedure followed in the design of the system. This procedure implies that employees are asked to

make the first move and take the initiative to define both their responsibilities and the performance

indicators for the measurement of those responsibilities. Not until the employees have come to an

agreement among themselves that adequate descriptions of their responsibilities and controllable per-

formance indicators have been developed are these responsibilities and indicators offered to the man-

agement in the form of a proposal to be discussed together. Following Maier’s (1973, p. 130)

statement: ‘People are more likely to accept the decisions that they feel responsible for, as a result

of having participated in making them,’ it is to be expected that carefully developed definitions of one’s

own responsibilities that are proposed to management for review and approval will have a personal

value or valence (Vroom, 1964) for those who have jointly developed these definitions. Thus, also

the motivational mechanisms mentioned in the participation literature as possible explanatory con-

cepts may be expected to play a role, given the form of participation examined in this study.

Together, the above arguments form the basis for our first hypothesis:

Hypothesis 1: Performance will improve after the implementation of a performance management

system that has been designed participatively in a form as prescribed by ProMES.

PARTICIPATION IN PERFORMANCE MANAGEMENT 833


As said before, the implementation of a participatively developed ProMES PMS entails a number of

interventions. We do not believe that the exclusion of just one of those interventions, i.e., the partici-

pative development of the PMS, will bereave the remaining ones completely of their effectiveness.

There are at least two reasons why some performance improvement might also be expected from a

tell-and-sell introduction of a performance management system identical to a system that has been

developed participatively. First, explanation and clarification of the arguments why the indicators cho-

sen indeed meet the controllability criterion may be expected to have some impact on the cognitive

mechanisms. Secondly, and perhaps more importantly, the actual implementation of a performance

management system provides employees with high-quality feedback on their performance and with

the possibility of discussing this feedback. On the basis of these latter interventions alone, some per-

formance improvement would be expected (Algera, 1990; Alvero et al., 2001; Balcazar et al., 1986;

Kluger & DeNisi, 1996; Locke & Latham, 1990, 2002). This leads to our second hypothesis:

Hypothesis 2: Performance will improve after the implementation of a performance management

system identical to the one mentioned in Hypothesis 1, when that system has been introduced by

means of a tell-and-sell strategy.

Although we expect at least some positive effects from the implementation of a high-quality PMS,

introduced in a tell-and-sell way, we do believe that there is added value in developing such a system

participatively. The main reason is that in a tell-and-sell introduction employees are only told that other

employees apparently have become convinced of the controllability of performance indicators through

a thorough discussion of this issue. This seems to be a far weaker intervention with a far lesser impact

on cognitions than developing performance indicators oneself, discussing the controllability of those

indicators, and having an impact on the decision of whether or not to include those indicators in the

final system. A weaker effect is expected from a tell-and-sell introduction of an identical performance

management system, because in the case of tell-and-sell there has been no personal investment by

those to whom the system is told and sold, in the definitions of the responsibilities and the indicators

included in that system. These arguments form the basis for our third hypothesis:

Hypothesis 3: Implementation of a participatively (in a form as prescribed by ProMES) designed

performance management system will result in a larger performance increase compared to the

implementation of an identical system preceded by a tell-and-sell introduction of that system.

Although we believe that the above reasoning provides good arguments in favor of our three hypoth-

eses, we are aware of the debate in the literature on whether one might expect participation to have any

effects on performance at all. What we have learned from this debate is that the question, when phrased

in terms as general as used here, is an unanswerable one. Already at the start of the debate (Locke &

Schweiger, 1979), it was not only concluded that there are numerous mechanisms both cognitive and

motivational through which participation in decision making (PDM) may produce high performance,

but also that the effectiveness of PDM depends on numerous contextual factors. The PDM research

conducted in laboratory settings has not succeeded in encompassing this complexity (Schweiger &

Leana, 1986, p. 159), whereas field studies have been struggling with disentangling participation

and other causal and contextual variables. The latter point was demonstrated by the reactions to the

work of Cotton and his colleagues, who added form of participation to the specter of potential causal

factors (Cotton et al., 1988; Cotton, Vollrath, Lengnick-Hall, & Froggatt, 1990). Based on a narrative

classification of participation studies, they argued that forms of participation that were formal, direct,

and long term, with high employee influence on an important issue (such as participation in work deci-

sions, consultative participation), were more effective than other forms of participation (such as infor-

mal participation and representative participation). In their reanalysis of Cotton et al.’s (1988) studies,



Leana, Locke, and Schweiger (1990) excluded multivariate studies involving manipulations beyond

participation that, in their opinion, could not be used to support participation as a separate variable.

Only 21 per cent of the remaining studies reported positive effect of PDM on performance (against 57

per cent of the studies in Cotton et al.’s, 1988, classification). However, the majority of studies that

remained involved forms of participation such as short-term participation and representative participa-

tion that were judged the least effective in Cotton et al.’s (1988) review. Thus, one could argue that the

failure to isolate the effects of participation in many of the studies involving the potentially most effec-

tive forms of participation impeded an effective test of the proposition that different forms of partici-

pation are associated with different outcomes.

We agree with Sagie (1994, 1995) that inconsistent findings on the effects of PDM on performance

ask for a search for moderators, which may be found among the contextual factors mentioned above.

We also agree with Wagner et al. (1997) that researchers should be more explicit about the details of

their studies on participation, because without specifying form of participation, the assumed mechan-

isms, and relevant contextual factors, it will be practically impossible to evaluate the outcomes of those

studies. More recently (Latham et al., 1994; Wagner et al., 1997), successful attempts have been made

to get a grasp on the motivational and cognitive mechanisms already proposed by Locke and

Schweiger in 1979. In line with the above suggestions, we present a study in which participation

focuses on the work roles of service engineers (taking into account several context factors with an

impact on controllability of work outcomes). The form of participation is formal (launched by man-

agement), long term (lasting over a period of almost 2 years), direct (immediate influence of all

employees to whom the issue is of concern), with a high degree of participant influence (decisions

being taken on the basis of participants’ inputs and the quality of their arguments), and about a sensi-

tive and important issue (the design of a ProMES system for the self-management of the participants’

performance). With reference to the assumed cognitive and motivational mechanisms (formulated by

NPI theory, expectancy-valence theory, and self-efficacy theory), we have also provided arguments

why this specific form of participation, given this specific context, should be more effective than

the tell-and-sell approach. The use of a quasi-experimental design in this realistic field context enables

an explicit test of the effectiveness of this form of participation.

Organizational Context

The Organization and its Environment

The organization was a Dutch supplier of office equipment, mainly photocopiers. The company did not

manufacture these machines, but bought them from leading Japanese suppliers, sold these under its own

name, and provided service. Since the 1970s, the office automation industry had moved from a product-

oriented industry via a price-oriented industry to an industry in which companies that were able to meet

the high demands placed on service had the competitive edge in a situation where product features and

prices hardly differentiated among photocopier suppliers. Through the use of effective and efficient

maintenance procedures and sophisticated planning systems, the company had been able to change

its image from ‘price fighter’ to all-round dependable supplier with after-sales service as its competitive

edge, thus securing a market share of approximately 15 per cent. The company’s customers ranged

from very small (e.g., private persons, small businesses) to very large (e.g., multinationals, universities,

municipalities). During the 1980s and early 1990s the company had grown from fewer than 100

employees to almost 650 employees, half of whom worked in the service department.



Method

Research design

The research design used to test the effects of the interventions was a quasi-experimental design with

control groups. First, the performance management system was participatively designed and imple-

mented in two of the 14 service regions, with the other 12 regions serving as the control condition.

After a baseline period and a period of feedback in these two regions, the system was introduced in

a tell-and-sell manner in six of the control regions, while the other six regions remained in the control

condition.

Development and implementation of the ProMES system

System development in the participation condition

The development of the ProMES system took 15 monthly meetings with both regions participating in

system development, including one meeting to achieve consensus between the two regions on the

The Field Service DepartmentThis department consisted of 14 geographically dispersed regions. Each service region was headed

by a supervisor and consisted of 15–20 service technicians, responsible for individually carrying out

preventive maintenance according to prescribed routines, and diagnosing and solving machine mal-

functions at clients’ offices. Service was provided on approximately 75 types of photocopiers sold

up to then, ranging from small copiers for infrequent use (‘low volume’), to large copiers with addi-

tional features for professional use (‘high volume’). This meant that the work-flow was highly com-

plex, since many different types of photocopiers had to be serviced and technical characteristics of

each type influenced the results that could be achieved. Depending on experience, technicians

worked on 2–10 types of photocopier.

The field service department was characterized by a lack of communication, both horizontally

(within regions) and vertically (between regions and service management). The departmental cul-

ture was ‘top down’: decisions were made at the head office, in most cases without consultation of

those affected. Projects initiated by management were usually greeted less than enthusiastically by

the work floor.

The management of the field service department had three reasons for starting a performance

management program for its service technicians. Firstly, it could help maintain or improve the qual-

ity of service while reducing its cost, securing the competitive edge. Secondly, it would meet the

technicians’ recurrent request for valid and useful feedback about their performance (until then,

performance feedback had consisted of one or two measures of repair quality, which were difficult

to interpret because they did not take into account the differences between types of copiers). The

third reason involved a possible future inclusion of performance information in the performance

appraisal of the technicians. The management agreed to a participative program, hoping that it

would lead to a high-quality system, that would be accepted by the employees.

The Time-Frame

The study was conducted in the early to mid 1990s.



design of the system. Limited possibilities for horizontal communication combined with the size of the

development teams necessitated the use of non-interactive information-gathering techniques (nominal

group technique and Delphi technique) to sustain the approximately 20 hours of group discussion to

reach consensus on the design of system components. Four of the meetings were review and approval

meetings in which the design teams presented elements of the system (key result areas and indicators,

contingencies, feedback report, final complete system) to the management and agreement was reached.

After 21 months, an operational system was available. The ProMES system developed by the two

regions in the participation condition consisted of four performance indicators covering two key result

areas (see Table 1).

The feedback provided to the technicians consisted of two monthly feedback reports: one report

covering a technician’s individual performance and one report covering the performance of his or

her region as a whole. In addition, graphical feedback on performance trends covering up to 12 months

was provided.

System implementation in the participation condition

Immediately after the system development was completed, the performance management system was

implemented in the two regions that had developed the system. Because the technicians of the parti-

cipation regions were well aware of the contents and purposes of the system developed by them, the

implementation phase was started with no further activities than an invitation to use the feedback as a

tool for performance improvement and to report experiences with and comments on the system. A

reference manual was provided with detailed information on the system, as were feedback reports cov-

ering a technician’s performance in the 3 months prior to implementation.

System introduction and implementation in the tell-and-sell condition

Nine months after the start of the implementation phase in the participation condition, the system was

introduced in the six regions that took part in the tell-and-sell condition. Through a standardized intro-

duction meeting, the technicians of each of the six regions were provided with a basic introduction into

the ProMES system that they were going to use, and they got an opportunity to ask questions (for

Table 1. Key result areas, performance measures, and importance weights

Key result Performance indicator Description Weight ofarea indicator

Quality Mean copies between calls Average number of photocopies the client’s machine 100has produced between two consecutive service calls

Percentage repeat calls Percentage of calls that occur within five working 70days from the previous call on the same machine

Cost Labor time per call Average time spent per service call on maintenance 80and repair activities (minutes)

Parts cost per call Average amount of spare parts used per service call 50(monetary value)

Note: The relative importance of the indicators was agreed upon by technicians, supervisors, and upper management, based on(1) their estimates of changes in customer satisfaction of a comparable performance change on each of the four indicators and (2)an activity-based costing analysis, which resulted in estimates of the financial consequences of a comparable performancechange on each of the four indicators. Consensus on the final weights was reached among the two regions’ technicians,supervisors, and upper management in a review and approval meeting. The composite measure of overall performance took intoaccount that a technician’s workload may change from one month to another (more copiers of type A in one month, more copiersof type B in another month) and that technicians did not work on the same types of photocopiers (one technician may primarilywork on type C, another technician on type D). This ensured comparability of performance across technicians.



further details see below). The same reference manual as above was provided to the technicians in the

tell-and-sell condition, as were feedback reports covering a technician’s performance in the 3 months

prior to the meeting.

Both experimental conditions

Parallel to the introduction of the system in the tell-and-sell condition, bilateral feedback meetings

between the supervisor and his technicians were set up in both experimental conditions. These meet-

ings were a response to the growing need of technicians in the participation condition for help to make

still more effective use of the feedback provided by the ProMES system. Technicians needed support

from their supervisors to use their feedback reports as problem-solving tools. To that end, a procedure

for individual feedback sessions was developed and in order to convey to the supervisors the knowl-

edge and skills to conduct effective bilateral feedback meetings with their technicians, a training pro-

gram, based on behavior modeling principles (e.g., Goldstein & Sorcher, 1974; Latham & Saari, 1979;

Latham & Wexley, 1981), was designed and carried out. The training program focused on a detailed

explanation of the workings of the system, and centered on ways to effectively introduce the system to

individual engineers and discuss their feedback reports. The categories of supervisor behaviors used as

key learning points were:

� achieving technicians’ understanding and acceptance: ‘Explaining the objectives and details of the

procedure’ and ‘Dealing with possible resistance;’

� setting technicians’ self-management in motion: ‘Analyzing feedback reports’ and ‘Looking for

improvement areas;’

� general behaviors: ‘Discussion until consensus’ and ‘Maintaining a constructive atmosphere.’

As the above supervisor training and concomitant feedback sessions, if only provided to supervisors

and technicians of the participation condition, would obviously have added difference between experi-

mental conditions over and above the one examined in this study, the supervisor training and feedback

sessions were also implemented in the tell-and-sell condition. In doing so, differences between experi-

mental conditions were restricted to the main factor of concern, namely form of participation in system

design.

Operationalization of variables

As the main dependent variable, performance data of individual technicians were used. In all three

conditions (n¼ 182, all men) these data were collected from the feedback reports generated by the

ProMES system. The overall ProMES performance index was used as the basis for evaluating

employee performance in all three conditions. Since the technicians were grouped into 14 regions,

we needed to determine whether the individual technician—in contrast to the region—would be

the appropriate unit of analysis. Applying the procedures and guidelines recommended by Muthen

and Satorra (1995) and Snijders and Bosker (1999) to the performance changes associated with imple-

mentation of the ProMES system in the participation and tell-and-sell condition, we found that both

conditions met their criterion for using individual data, i.e., a design effect of 2 or less.

The independent variable, degree of participation, had three levels, namely participation, tell-and-

sell, and control group. Participation as operationalized by us implied the following. First, a group of

technicians, together with their supervisor, through a process of discussion until consensus generates

definitions of the technicians’ result areas and performance indicators for the measurement of those

results, indications of the relative importance of the performance indicators, and the basic elements

of a feedback system through which the technicians will be informed on their performance on the



indicators. Secondly, the group (including its supervisor) reaches agreement, again through a process

of discussion until consensus, on the above issues with the management responsible for the group. In

both phases of the process, a facilitator takes care of a number of discussion rules (e.g., decisions have

to be based on arguments, not on authority; performance indicators have to fulfill specific design

requirements, the most important one being controllability).

Tell-and-sell as operationalized by us implied the following. A meeting was held, which included:

� an opening statement by the service manager, stressing the importance of performance improvement

for the company and their support for the ProMES system as a tool for achieving this;

� an introduction into performance management (e.g., criteria for good indicators, feedback, and

goals) and the ProMES method as used by the two development regions;

� an explanation of the performance management system as developed by colleagues from the two

participation regions;

� an opportunity for questions and answers;

� an invitation to use the feedback as a tool for performance improvement and to report experiences

with and comments on the system.

In the control group, performance was measured in the same way as in the participation and tell-and-

sell groups. ProMES feedback was not provided.

Equating the conditions

Care was taken to obtain conditions that were equivalent on relevant variables. Retrospectively recon-

structed regional performance data for the ProMES indicators developed during the study showed that

the average pre-study performance of the participation condition (two regions) and control condition

(12 regions) was the same. Furthermore, the tell-and-sell condition (six regions) and remaining control

condition (six regions) were composed in such a way that they had the same overall ProMES score in

the period before the start of the tell-and-sell implementation. The mean age of the technicians and

their tenure were almost identical in all three conditions (Age—participation: 27.2 (SD¼ 4.1); tell-

and-sell: 26.4 (SD¼ 4.0); control: 26.7 (SD¼ 4.2). Tenure—participation: 3.6 (SD¼ 2.8); tell-and-

sell: 3.6 (SD¼ 3.2); control: 3.7 (SD¼ 3.2). All data refer to the first month of the baseline). Lastly,

the ratio of technicians working in urban versus rural areas—with possibly different customer

demands—was practically the same in all three conditions (participation: 48 per cent urban; tell-

and-sell: 45 per cent urban; control: 44 per cent urban).

Additional measures

In the participation condition, the perceived amount of participation was assessed as a manipulation

check with two items: ‘In the development of the ProMES system, the technicians’ opinions were suf-

ficiently taken into account’ (five-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly

agree)) and ‘How much influence did the technicians have on the final ProMES system?’ (percentage

from 0 to 100).

In the implementation of ProMES systems, feedback is the primary intervention employed to

achieve performance improvement (e.g., Pritchard et al., 1988, 1989). We included two variables to

measure the impact of our intervention with regard to feedback in the three conditions. ‘Amount of

feedback’ assessed whether the treatments in the two experimental conditions actually had any signif-

icant impact on the technicians’ daily work (manipulation check). It is expected to be judged signifi-

cantly higher in the experimental conditions compared to the control condition. ‘Usefulness of

feedback’ was a critical variable used to explain performance differences among the three conditions.



Although formal goal-setting is only a secondary intervention in ProMES implementations (e.g.,

Pritchard, 1995a), and had only received limited attention in our supervisor training program, implicit

or spontaneous goal-setting could have occurred as a result of the feedback meetings. Measures of goal

clarity and goal difficulty were therefore included. Variables related to supervisor behaviors (‘Influ-

ence in goal-setting,’ ‘Achieving understanding and acceptance,’ ‘Feedback’) were employed as con-

trol variables to rule out alternative explanations (to participation in system design) for any

performance differences found. Thus, at the end of the study, a questionnaire on feedback and goal-

setting was administered in all three conditions. It contained the following seven variables: amount of

feedback (manipulation check; 4 items; sample item: ‘I receive information about my performance’),

usefulness of feedback (process variable; 5 items; sample item: ‘From the information I receive about

my performance, I can easily gather how well I’m doing’), clarity of quality goals (process variable; 2

items; sample item: ‘Precise agreements are made about the level of quality I have to attain’), clarity of

cost goals (process variable; 2 items; sample item: ‘Precise agreements are made about my use of spare

parts’), influence in goal-setting (control variable; 3 items; sample item: ‘When agreements are made

about my work performance, my opinion is taken into account’), goal difficulty (process variable; 4

items; sample item: ‘I think it is rather difficult to meet the requirements for use of spare parts’),

knowledge of priorities (process variable; 4 items; sample item: ‘I know exactly which aspects of

my work are very important and which are not so important’). For each item, a five-point Likert scale

was used, ranging from 1 (almost never) to 5 (almost always). All items and scales were based on

Algera and Van Tuijl (1990). A confirmatory factor analysis, using LISREL 8.53, yielded a reasonable

fit for the predicted seven-factor model (RSMEA¼ 0.083, CFI¼ 0.92). Additionally, in the participa-

tion and tell-and-sell condition, a second questionnaire was used to measure the supervisor behaviors

and the technicians’ satisfaction with the feedback meetings. The following scales were developed for

the measurement of supervisor behaviors: achieving understanding and acceptance (control variable; 4

items; sample item: ‘My supervisor paid serious attention to my opinion about ProMES and any objec-

tions I had’), feedback (control variable; 4 items; sample item: ‘In discussing my feedback reports, my

supervisor and I looked for causes of high scores and for causes of low scores’), satisfaction with the

ProMES program (process variable; 2 items; sample item: ‘The ProMES meetings are a useful way of

working on improving my performance’). For each item, a five-point Likert scale was used, ranging

from 1 (strongly disagree) to 5 (strongly agree). Table 2 presents descriptive statistics and correlations

for all variables.

Results

The presentation of the results follows the order of system implementation in the experiment. The

results are summarized in Table 3.

First implementation phase: implementing the system in the regions thatdeveloped the system participatively (months 9–18)

Performance data for all individual technicians in the study were available (retrospectively) for the last

8 months of system development. These 8 months (months 1–8) were treated as the baseline period.

The 10-month period of system implementation in the participation condition was associated with a

significant increase of overall performance. The participation condition (n¼ 31) attained an average



Tab

le2

.A

dd

itio

nal

mea

sure

s:d

escr

ipti

ve

stat

isti

cs,

corr

elat

ion

s,an

dre

liab

ilit

iesa

NM

SD

12

34

56

78

91

0

1A

mo

un

to

ffe

edb

ack

17

53

.44

1.1

4(0

.91

)2

Use

fuln

ess

of

feed

bac

k1

75

3.0

30

.96

0.6

5*

*(0

.90

)3

Cla

rity

of

qu

alit

yg

oal

s1

75

3.5

80

.71

0.0

00

.21

**

(0.7

8)

4C

lari

tyo

fco

stg

oal

s1

75

2.9

60

.85

0.1

20

.27

**

0.4

8*

*(0

.74

)5

Go

ald

iffi

cult

y1

75

2.3

30

.89

0.1

1�

0.0

5�

0.2

6*

*�

0.3

1*

*(0

.87

)6

Infl

uen

cein

go

al-s

etti

ng

17

53

.39

0.8

40

.16

*0

.26

**

0.3

4*

*0

.40

**

�0

.21

**

(0.7

7)

7K

now

led

ge

of

pri

ori

ties

17

54

.03

0.6

70

.07

0.1

40

.27

**

0.2

9*

*�

0.2

8*

*0

.26

**

(0.7

7)

8S

up

erv

iso

rb

ehav

iors

10

03

.85

0.7

40

.11

0.2

2*

0.2

1*

0.2

2*

�0

.06

0.4

2*

*0

.27

**

(0.7

2)

‘Ach

iev

ing

un

der

stan

din

gan

dac

cep

tan

ce’

9S

up

erv

iso

rb

ehav

iors

56

3.3

71

.01

0.2

30

.23

0.2

9*

0.2

50

.00

0.3

0*

*0

.14

0.4

8*

*(0

.78

)‘F

eed

bac

k’

10

Sat

isfa

ctio

nw

ith

10

03

.26

0.9

10

.29

**

0.2

9*

*0

.06

0.2

6*

�0

.14

0.3

1*

*0

.08

0.1

50

.45

**

(0.7

3)

the

Pro

ME

Sp

rog

ram

aV

aria

ble

s1

–7

wer

em

easu

red

inal

lth

ree

con

dit

ion

s.V

aria

ble

s8

–1

0w

ere

mea

sure

din

the

par

tici

pat

ion

and

tell

-an

d-s

ell

con

dit

ion.

Th

eref

ore

,n

ot

all

corr

elat

ion

coef

fici

ents

are

bas

edo

nth

esa

me

n.

Th

ed

iag

on

alco

nta

ins

the

Cro

nb

ach�

ind

exo

fin

tern

alco

nsi

sten

cy.

*p<

0.0

5;

**p<

0.0

1.



19-point increase in the ProMES effectiveness score compared to the control condition (n¼ 151) (see

Figure 1). A 2� 2 multivariate analysis of variance (MANOVA), in which the mean overall score from

the baseline period and the feedback period are considered repeated measures, revealed a significant

interaction effect (F¼ 11.44, p< 0.001). These results offer support for the first hypothesis: perfor-

mance improved significantly after implementation of the participatively designed performance man-

agement system.

Second implementation phase: implementing the system in the regionsthat got a tell-and-sell introduction into the system developed bythe participation regions (months 19–27)

The 9-month period of system implementation in the tell-and-sell condition (n¼ 80) resulted in a 12-

point increase of the overall ProMES effectiveness score compared to the control condition (n¼ 71), as

shown in Figure 2. A 2� 2 multivariate analysis of variance (MANOVA), in which the mean overall

Table 3. Performance changes in the participation condition and the tell-and-sell condition relative to controlcondition

First implementation Second implementation Mean overall Comparison ofphase (months 9–18) phase (months 19–27) performance change the performance

relative to the baseline relative to the previous phase relative to the baseline changes

Participation Overall þ19 Overall: þ8 Overall: þ26(n¼ 31) F¼ 11.44, p< 0.001a F¼ 1.49, n.s. F¼ 19.47, p< 0.001a

F¼ 4.65, p< 0.05c

Tell-and-sell (Baseline) Overall: þ12 Overall: þ12(n¼ 80) F¼ 5.81, p< 0.05b F¼ 5.81, p< 0.05b

aSupports Hypothesis 1: performance improved after implementation of the performance management system in the participationcondition.bSupports Hypothesis 2: performance improved after implementation of the performance management system in the tell-and-sellcondition.cSupports Hypothesis 3: implementation of the performance management system was associated with a larger performanceimprovement in the participation condition, compared to the tell-and-sell implementation.

Figure 1. Effects of implementing ProMES in the participation condition (months 9–18). The vertical axisrepresents the average monthly overall effectiveness value obtained in the two conditions



score from the baseline period (months 9–18) and the feedback period (months 19–27) are considered

repeated measures, showed that this increase was significant (F¼ 5.81, p< 0.05).

These results support the second hypothesis: performance improved significantly after implementa-

tion of the performance management system that was introduced in a tell-and-sell manner.

Second implementation phase: continuation of feedback in the regionsthat developed the system participatively (months 19–27)

Adding the bilateral feedback meetings to the existing feedback in the participation condition (n¼ 31)

resulted in an additional eight-point increase compared to the control condition (n¼ 71), also shown in

Figure 2. This increase was not significant (F¼ 1.49, n.s.). The mean overall increase (months 9–27) in

the participation condition (n¼ 31) relative to the control condition (n¼ 71) amounted to 26 points. A

2� 2 MANOVA, in which the mean overall score from the baseline period (months 1–8) and the feed-

back period (months 9–27) are considered repeated measures, revealed that this increase was signifi-

cant (F¼ 19.47, p< 0.001). These results also support the first hypothesis: performance improved

significantly after implementation of the participatively designed performance management system.

Second implementation phase: comparing the effects of system implementationin the participation and tell-and-sell regions

The overall effects in the participation and tell-and-sell condition were compared as follows. First, the

mean overall baseline score relative to the mean control group score was calculated for the participa-

tion and tell-and-sell condition. Analogously, for both groups the mean overall score in the feedback

period relative to the mean control group score was calculated (for the participation groups the

Figure 2. Effects of implementing ProMES in the tell-and-sell condition and of continued feedback in theparticipation condition (months 19–27). The vertical axis represents the average monthly overall effectiveness

value obtained in the three conditions



feedback period covered months 9–27, for the tell-and-sell group the feedback period covered months

19–27). A 2� 2 MANOVA, in which these baseline and feedback scores (relative to the control group)

are considered repeated measures for the participation and tell-and-sell group, produced a significant

interaction effect (F¼ 4.65, p< 0.05). This result supports the third hypothesis: implementation of a

participatively designed performance management system was associated with a larger performance

improvement in the regions that developed the system, compared to the regions in which the same

system was implemented after a tell-and-sell introduction.

Additional measures

With regard to participation in system development, the technicians in the participation condition

agreed with the statement that their opinion had been sufficiently taken into account (4.3 on a five-

point Likert scale). Also, on average they reported a 69 per cent (SD¼ 16 per cent) influence on

the contents of the system. These results confirm that these technicians were allowed (and exerted)

a high degree of influence on the final system.

In Table 4, the results of ANOVA and post hoc tests of differences are shown for the variables related

to feedback and goals that were measured in all three conditions.

As would be expected, the technicians in the participation and tell-and-sell condition reported a sig-

nificantly higher amount of feedback than their colleagues in the control condition (t(94)¼ 11.00,

p< 0.001, and t(144)¼ 11.62, p< 0.001, respectively). The three conditions differed significantly

with regard to perceived usefulness of feedback. Both in the participation and in the tell-and-sell con-

dition, perceived usefulness of feedback was higher than in the control condition (t(94)¼ 7.08,

p< 0.001, and t(144)¼ 4.84, p< 0.001, respectively). Participants in the participation condition con-

sidered the feedback significantly more useful than those in the tell-and-sell condition (t(106)¼ 2.58,

p< 0.05). The only other significant result was a higher perceived goal difficulty in the tell-and-sell

condition compared to the control condition (t(144)¼ 3.53, p< 0.001). The three conditions did not

differ significantly with regard to ratings of clarity of quality goals, clarity of cost goals, influence in

goal-setting, and knowledge of priorities.

Table 4. Additional measures: results of one-way analysis of variance, comparing the participation, tell-and-sell,and control condition

Condition: Participation Tell-and-sell Control ANOVAResponse: n¼ 29 (94%) n¼ 79 (99%) n¼ 67 (94%) F(2, 172)

1 Amount of feedbackb,c 4.23 (0.61) 4.01 (0.86) 2.41 (0.80) 90.3**2 Usefulness of feedbacka,b,c 3.71 (0.63) 3.22 (0.95) 2.51 (0.82) 23.9**3 Clarity of quality goals 3.50 (0.64) 3.49 (0.80) 3.72 (0.59) 2.34 Clarity of cost goals 2.84 (0.75) 2.87 (0.99) 3.13 (0.67) 2.05 Goal difficultyc 2.31 (0.84) 2.57 (0.97) 2.06 (0.73) 6.3*6 Influence on goal-setting 3.24 (0.85) 3.39 (83) 3.45 (0.84) 0.67 Knowledge of priorities 3.93 (0.84) 4.07 (0.61) 4.01 (0.65) 0.5

aMultiple comparison (Scheffe) tests yielded statistically significant differences (p< 0.05) for the participation and the tell-and-sell condition.bMultiple comparison (Scheffe) tests yielded statistically significant differences (p< 0.001) for the participation and the controlcondition.cMultiple comparison (Scheffe) tests yielded statistically significant differences (p< 0.001) for the tell-and-sell and the controlcondition.*p< 0.05; **p< 0.001.



Table 5 contains the results for the supervisor behaviors with respect to ‘Achieving

understanding and acceptance’ and ‘Feedback’ and the technicians’ overall satisfaction with the

ProMES program.

The supervisors in both conditions implemented the key behaviors for both ‘Achieving understand-

ing and acceptance’ and ‘Feedback’ to a reasonable degree. There was no significant difference

between the participation and the tell-and-sell condition for these variables (t(98)¼�1.58, n.s. and

t(98)¼�0.42, n.s., respectively), which suggests that performance differences between the participa-

tion and tell-and-sell condition were not caused by differences in the effectiveness of supervisor beha-

viors (if the performance effects had been due to differential supervisor behaviors, these should have

been significantly higher in the participation condition). Satisfaction with the ProMES program was

significantly higher in the participation condition compared to the tell-and-sell condition (t(98)¼ 2.41,

p< 0.05).

Discussion

This study has shown that implementing a participatively designed performance management system

can lead to a significantly larger performance increase, compared to the performance increase that fol-

lows the tell-and-sell introduction of an identical system. In addition, in both experimental conditions a

significant performance increase compared to the control condition has been demonstrated. Thus, all

three hypotheses have been confirmed, thereby providing compelling evidence for our main tenet that

participation can make a difference compared to tell-and-sell, depending on the circumstances (e.g.,

the importance of the issue and the kind of setting). The effect sizes found further sustain this position.

We compared the effect sizes in the participation and tell-and-sell condition with those found by

Wagner (1994) in his meta-analysis of participation studies. To calculate the effect sizes for our study,

we entered the F-values from Table 3 into the equations provided by Hunter and Schmidt (1990) and

Rosenthal (1991). This resulted in d¼ 0.96 for the participation condition and d¼ 0.39 for the tell-and-

sell condition. Wagner (1994) reported correlations between r¼ 0.15 and r¼ 0.25 between participa-

tion and performance (excluding percept–percept correlations), which is equivalent to effect sizes

between d¼ 0.30 and d¼ 0.48 (cf. Rosenthal, 1991). Thus, the participation effect size in our study

clearly exceeds the upper limit of the range of effect sizes reported by Wagner (1994), while the tell-

and-sell effect size lies within that range. This at least demonstrates that the impact of participation

Table 5. Additional measures: supervisor behaviors and satisfaction with the program

Condition: Participation Tell-and-sell tResponse: n¼ 27 (87%) n¼ 73 (91%)

8 Supervisor behaviors ‘Achieving 3.64 (0.73) 3.91 (0.73) �1.58 (d.f.¼ 98)understanding and acceptance’

9 Supervisor behaviors ‘Feedback’a 3.28 (1.17) n¼ 16 3.41 (0.95) n¼ 40 �0.42 (d.f.¼ 54)10 Satisfaction with the ProMES program 3.61 (0.74) 3.13 (0.94) 2.41* (d.f.¼ 98)

aAlthough almost every technician went through an introductory meeting, actual discussion of the feedback reports occurred with56% of the technicians (59% participation, 55% tell-and-sell). This explains the lower n for this variable. Almost no formalquantitative goal-setting occurred (with one technician (3%) in the participation condition and with five technicians (7%) in thetell-and-sell condition).*p< 0.05.



when it comes to issues and settings as investigated in our study might be much larger than suggested

by researchers who are skeptical with regard to the added value of participation (see, for example,

Locke & Schweiger, 1979; Schweiger & Leana, 1986; Wagner, 1994).

With respect to the cognitive and motivational mechanisms mentioned in the introduction, we would

like to further speculate on the causes of the performance difference found between the participative

condition and the tell-and-sell condition. A closer examination of the performance effects in both con-

ditions revealed that feedback in the participation condition led to improved overall performance

mainly through improvements on the cost indicators (Quality þ7, Cost þ19, Overall þ26), while in

the tell-and-sell condition improvement on the quality indicators contributed most to overall perfor-

mance increase (Quality þ10, Cost þ2, Overall þ12). Thus, while both conditions improved almost

equally in terms of quality performance, the participative condition in addition improved considerably

in terms of cost performance, whereas the tell-and-sell condition hardly improved on cost at all. It

should be noted that before the start of the project a generally held belief among service technicians

was that ‘a good technician is one who attains high quality.’ The discussions with management during

system development in the participative condition brought to light that cost performance was almost

equally important for the company as quality performance. Therefore, one may conclude that the per-

formance feedback in the tell-and-sell condition had its effect through building forth on the long-held

belief in the importance of quality. In the participation condition, on the other hand, this long-held

belief was effectively challenged during the discussions and changed into a more balanced approach

to quality and cost performance. Thus, a plausible explanation is that the technicians in the participa-

tion condition had become committed to the priorities that they—in agreement with the manage-

ment—had established, whereas the tell-and-sell intervention was not powerful enough to

effectively challenge the long-held belief among service technicians that ‘a good technician goes

for quality.’ In this case, participation appears to have had an important motivational effect, through

the willingness to allocate effort to cost aspects of the task. This explanation is supported by the results

on two of the additional measures, which were in line with our predictions. Whereas the perceived

amount of feedback was the same in the participation and tell-and-sell condition (both significantly

higher than the control condition), the usefulness of the feedback was judged significantly higher in

the participation condition. Also, the usefulness of the bilateral feedback meetings for attaining per-

formance improvement was rated significantly higher in the participation condition. Apparently,

acceptance of the feedback on both quality and cost performance, as well as acceptance of attempts

by the organization to stimulate optimization of overall performance through joint optimization of

both quality and cost performance, was higher in the participation condition than in the tell-and-sell

condition. Technicians in the participation condition will more likely have followed through on these

priorities (and the apparent opportunities for improvements on cost indicators) and, thus, have attained

a larger overall performance increase. Alternative explanations involving differences in the effective-

ness of supervisor behaviors and/or perceived influence in goal-setting are less likely since the parti-

cipation and tell-and-sell condition did not differ on any of these measures.

We had expected to find differences between the experimental conditions and the control condition

on the measures of clarity of goals, influence in goal-setting, and knowledge of priorities. There are

several possible explanations for the absence of differences on these variables. First, one should realize

that the main characteristic of a ProMES system is that it provides feedback. It can be used as a goal-

setting system too, but this has in most of the cases known to us only been done after a longer period of

use as a feedback system. In the present study (notwithstanding the presence of a goal-setting module

in the training program that was provided), setting specific and challenging goals played a negligible

role. This may explain why participants in the three conditions did not react differently with regard to

the questionnaire items on goals. A second explanation might be that participants in different condi-

tions were using different frames of reference in interpreting questionnaire items (cf. Golembiewski,



Billingsley, & Yeager, 1976). For example, it became apparent that ‘clarity of cost goals’ had been

interpreted by some technicians in the control condition as the degree to which the maintenance pro-

cedure guidelines for spare part replacement and time expenditure were clear, whereas the technicians

in the experimental conditions mainly referred to agreements with regard to the operationalization of

the cost indicators.

One final comparison between the participation condition and the tell-and-sell condition may be of

interest. We found a higher perceived goal difficulty in the tell-and-sell condition combined with no

difference on the knowledge of priorities measure. Taking into account the limited role of explicit goal-

setting, this perhaps indicates that at least some of those in the tell-and-sell condition—although aware

of what was expected—were less certain how to fulfill these expectations. A tentative explanation may

be that cognitive benefits with regard to effective task strategies ensued from participatively develop-

ing the components of the ProMES system.

Limitations

In this field study, we were able to employ a relatively strong research design: quasi-experimental with

control groups. The groups were equated on several characteristics considered relevant, such as pre-

intervention performance. Furthermore, the results for the additional measures show that the different

effects of participation and tell-and-sell on performance cannot be explained by differences in the

amount of feedback or differences with regard to supervisor behaviors. Nevertheless, the study has

some limitations, which need to be mentioned. In the introduction, we have argued that the form of

participation we addressed can only be studied in a real-life organizational environment with its inher-

ent complexity. Even though we employed a quasi-experimental design and made every effort possible

to eliminate threats to internal validity, incomplete control of the research environment precludes rul-

ing out all possible rival explanations. Nevertheless, we feel confident that we have countered most

threats to the internal validity of the study. The fact that the conditions were equal with respect to

pre-intervention performance as well as on other variables will have contributed to prevent testing,

statistical regression, history, and differential selection effects from occurring. The maturation threat

was countered through the use of a control condition. Also, the use of identical performance measures

in all conditions during the entire study practically ruled out instrumentation as an alternative explana-

tion for our results. Furthermore, experimental mortality could be ruled out since turnover was very

low (e.g., only seven technicians were excluded from the study because they left the company during

the feedback phase). However, other threats cannot be ruled out completely, although it is not likely

that they played a significant role. For example, it is conceivable that some participants in the study

might have tried to ‘look good in the figures’ by manipulating the measures. In our study, the data on

which the performance measures were based were reported by the technicians themselves (e.g., copy

counters, spare parts used). However, the objective nature of the performance measures plus the fact

that incorrect data supplied by a technician would most likely have been spotted by a colleague, his

supervisor, or by anyone performing checks and balances on available data, would render systematic

fraud unlikely (which would then present a testing threat if it occurred to a different degree in the three

conditions). A multiple treatment inference threat might have occurred if, for example, the participa-

tion regions had been chosen based on experiment proneness or on a favorable leadership style of the

supervisors. Since this was not the case, occurrence of this threat is unlikely. Spill-over effects cannot

be completely discounted. However, because of the geographically dispersed regions, communication

between technicians from different regions was limited. We also have no indications that the lack of

opportunity to participate in system development caused negative feelings or reduced motivation in the

tell-and-sell or control condition.



Because of certain restrictions in the company due to which the supervisor training program had to

be postponed, the bilateral feedback meetings in the participation condition had to be preceded by sev-

eral months of feedback ‘through the mail’ (without feedback meetings). The feedback phase in the

tell-and-sell condition immediately started with the bilateral feedback meetings. Thus, in this respect,

the implementations in both experimental conditions were not completely identical. However, since

we used the mean performance increase to evaluate performance differences between the two condi-

tions, this difference in implementation would have run counter to our first and third hypotheses, since

it would have diminished the mean effect in the participation condition.

Our choice to use a tell-and-sell strategy in the non-participative condition and no ‘tell’

strategy might be seen as a limitation, although of a different kind than those above. Our choice reflects

the view that it is indeed a tell-and-sell strategy rather than a ‘tell’ strategy that should be investigated

as an alternative to participation in modern organizations. The real-life field situations that are

the frame of reference for the form of participation studied here are characterized by complex, mean-

ingful decisions within an organizational context, which have long-term consequences for employees.

Here, understanding and acceptance of the decision (and its consequences) by employees is a

major issue. Whereas it would be logical for a laboratory study (voluntary, short-term, context-free,

concerning an issue that is not very meaningful to the participants, once outside the laboratory)

to incorporate a ‘tell’ condition, this is hardly feasible in field studies such as the one we have con-

ducted. We think that it will be hard to find line managers or human resource managers who would

allow a ‘tell’ condition to be incorporated in any meaningful performance management study within

their company.

Implications and future research

This study has demonstrated that participation in the design of performance management systems can

be associated with unusually large increases in performance. The results of this study also imply that

the widespread view that a tell-and-sell persuasion strategy may serve as a substitute for participation

(e.g., Latham, Erez, & Locke, 1988; Sagie, 1994) might not apply to certain forms of participation in

field settings. Our study indicates that participation which is formal, direct, long-term, and applied

to all stages of a complex and meaningful decision within a real-life organizational setting may

be preferable to a tell-and-sell strategy. Future research could focus on determining which of these

aspects are—individually or in conjunction—responsible for the superiority of participation over

tell-and-sell.

In the introduction, we touched upon mechanisms that may explain the difference in performance

increase associated with participation versus tell-and-sell in the design and implementation of a per-

formance management system. In our study, we mainly focused on performance, the outcome variable.

Future research could add to our understanding by focusing on process variables, such as the cognitive

and motivational mechanisms through which participation and tell-and-sell are supposed to affect

performance.

This study also has implications for practitioners in the field of performance management. When

developing and implementing a performance management system as it was done in this study, one

may expect it to be less successful in the tell-and-sell units. Since having a full participation procedure

in each and every unit is hardly feasible, one would either have to accept this or employ a different

procedure, such as having representatives from each unit participate in all stages of system develop-

ment or having each unit participate in developing some part of the system (e.g., a key result area or an

indicator). Field studies comparing these approaches would contribute to our understanding of the

effects of different degrees of participation or different approaches to participation.



Acknowledgements

The authors thank Michael West for his helpful comments on a previous version of this manuscript, and

Gary Johns and three anonymous reviewers for their guidance in developing this manuscript.

Author biographies

Ad Kleingeld has a PhD in Industrial Engineering and Management Science from Technische

Universiteit Eindhoven in The Netherlands. He is assistant professor at the Department of

Technology Management, subdepartment of Human Performance Management, of this

university. His current research interests are performance management of interdependent individuals

and groups, task strategies, and the relationship between individuals’ goal orientation and their use of

feedback.

Harrie van Tuijl has a PhD in Experimental Psychology from Nijmegen University. He is Associate

Professor of Personnel Management at Technische Universiteit Eindhoven, Department of Technology

Management, subdepartment of Human Performance Management. His main research interests are:

productivity enhancement, organizational learning, group problem solving, task strategies, and co-

operation and competition.

Jen A. Algera has a PhD in social sciences from Leiden University. He is Professor of Personnel Man-

agement at Technische Universiteit Eindhoven, The Netherlands. His current work is on goal-setting

and feedback in combination with self-management of individuals and groups to arrive at high orga-

nizational productivity, especially in new technology environments.

References

Algera, J. A. (1990). Feedback systems in organizations. In C. L. Cooper, & I. T. Robertson (Eds.), Internationalreview of industrial and organizational psychology (Vol. 5, pp. 169–193). Chichester: Wiley.

Algera, J. A., & Van Tuijl, H. F. J. M. (1990). Feedback systems and the management of human performance inorganizations. In P. J. D. Drenth, J. A. Sergeant, & R. J. Takens (Eds.), European perspectives in psychology(Vol. 3, pp. 39–54). Chichester: Wiley.

Alvero, A. M., Bucklin, B. R., & Austin, J. (2001). An objective review of the effectiveness and essentialcharacteristics of performance feedback in organizational settings. Journal of Organizational BehaviorManagement, 21, 3–29.

Balcazar, F., Hopkins, B. L., & Suarez, Y. (1986). A critical, objective review of performance feedback. Journal ofOrganizational Behavior Management, 7, 65–89.

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman.Coch, L., & French, J. R. P. (1948). Overcoming resistance to change. Human Relations, 1, 512–532.Cotton, J. L., Vollrath, D. A., Froggatt, K. L., Lengnick-Hall, M. L., & Jennings, K. R. (1988). Employee

participation: diverse forms and different outcomes. Academy of Management Journal, 13, 8–22.Cotton, J. L., Vollrath, D. A., Lengnick-Hall, M. L., & Froggatt, K. L. (1990). Fact: the form of participation

does matter—a rebuttal to Leana, Locke, and Schweiger. Academy of Management Review, 15, 147–153.



Drucker, P. F. (1976). What results should you expect? A user’s guide to MBO. Public Administration Review, 36,12–19.

Goldstein, A. P., & Sorcher, M. (1974). Changing supervisor behavior. London: Pergamon Press.Golembiewski, R. T., Billingsley, K., & Yeager, S. (1976). Measuring change and persistence in human behavior:

types of change generated by OD designs. Journal of Applied Behavioral Science, 12, 133–157.Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: Correcting error and bias in research findings.

Newbury Park, CA: Sage.Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance: a historical review, a

meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin, 199, 254–284.Latham, G. P., Erez, M., & Locke, E. A. (1988). Resolving scientific disputes by the joint design of crucial

experiments by the antagonists: application to the Erez–Latham dispute regarding participation in goal setting.Journal of Applied Psychology, 73, 753–772.

Latham, G. P., & Saari, L. M. (1979). Application of social learning theory to training supervisors throughbehavior modeling. Journal of Applied Psychology, 64, 239–246.

Latham, G. P., & Wexley, K. N. (1981). Increasing productivity through performance appraisal. Reading, MA:Addison-Wesley.

Latham, G. P., Winters, D. C., & Locke, E. A. (1994). Cognitive and motivational effects of participation: amediator study. Journal of Organizational Behavior, 15, 49–63.

Leana, C. R., Locke, E. A., & Schweiger, D. M. (1990). Fact and fiction in analyzing research on participativedecision making: a critique of Cotton, Vollrath, Lengnick-Hall and Jennings. Academy of Management Review,15, 137–146.

Lewin, K. (1952). Group decision and social change. In G. E. Swanson, T. M. Newcomb, & E. L. Hartley (Eds.),Readings in social psychology (pp. 463–473). New York: Holt, Rinehart & Winston.

Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task performance. Englewood Cliffs, NJ:Prentice-Hall.

Locke, E. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation.American Psychologist, 57, 705–717.

Locke, E. A., & Schweiger, D. M. (1979). Participation in decision making: one more look. In B. Staw (Ed.),Research in organizational behavior (Vol. 1, pp. 265–339). Greenwich, CT: JAI Press.

Maier, N. R. F. (1973). Psychology in industrial organizations (4th ed.). Boston, MA: Houghton Mifflin.Muthen, B. O., & Satorra, A. (1995). Technical aspects of Muthen’s LISCOMP approach to estimation of latent

variable relations with a comprehensive measurement model. Psychometrica, 60, 489–503.Naylor, J. C., Pritchard, R. D., & Ilgen, D. R. (1980). A theory of behavior in organizations. New York: Academic

Press.Pritchard, R. D. (1990). Measuring and improving organizational productivity: A practical guide. New York:

Praeger.Pritchard, R. D. (Ed.). (1995a). Productivity measurement and improvement: Organizational case studies. New

York: Praeger.Pritchard, R. D. (1995b). Lessons learned about ProMES. In R. D. Pritchard (Ed.), Productivity measurement and

improvement: Organizational case studies (pp. 325–365). New York: Praeger.Pritchard, R. D., Holling, H., Lammers, F., & Clark, B. D. (Eds.). (2002). Improving organizational performance

with the productivity measurement and enhancement system: An international collaboration. New York: NovaScience.

Pritchard, R. D., Jones, S. E., Roth, P. L., Stuebing, K. K., & Ekeberg, S. E. (1988). The effects of feedback, goalsetting and incentives on organizational productivity. Journal of Applied Psychology, 73, 337–358.

Pritchard, R. D., Jones, S. E., Roth, P. L., Stuebing, K. K., & Ekeberg, S. E. (1989). The evaluation of an integratedapproach to measuring organizational productivity. Personnel Psychology, 42, 69–115.

Rodgers, R., & Hunter, J. E. (1991). Impact of management by objectives on organizational productivity. Journalof Applied Psychology, 76, 322–336.

Rosenthal, R. (1991). Meta-analytic procedures for social research. London: Sage.Sagie, A. (1994). Participative decision making and performance: a moderator analysis. Journal of Applied

Behavioral Science, 30, 227–246.Sagie, A. 1995. Employee participation and work outcomes: an end to the dispute? Academy of Management

Review, 20, 278–280.Schweiger, D. M., & Leana, C. R. (1986). Participation in decision making. In E. A. Locke (Ed.), Generalizing

from laboratory to field settings (pp. 147–166). Lexington, MA: Lexington Books.



Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel analysis: An introduction to basic and advanced multilevelmodeling. London: Sage.

Vroom, V. H. (1964). Work and motivation. New York: Wiley.Wagner, J. A. (1994). Participation’s effects on performance and satisfaction: a reconsideration of the evidence.

Academy of Management Review, 19, 312–330.Wagner, J. A., Leana, C. R., Locke, E. A., & Schweiger, D. M. (1997). Cognitive and motivational frameworks

in U.S. research on participation: a meta-analysis of primary effects. Journal of Organizational Behavior, 18,49–65.



participation in the design of performance management ... · the speciﬁc form of participation...

Documents